Moonshot AI releases Kimi K3.1 Pro with major score jump

Moonshot AI has released Kimi K3.1 Pro, a reasoning model that now holds the fourth spot on the Artificial Analysis Intelligence Index.

The new release scored 55 points. This result marks an 11 point gain over the previous Kimi K3 model. It sits just three points behind the leading proprietary systems.

K3.1 Pro is a post-trained version of K3. It used significantly more compute during that process. Only three models currently score higher. They are Anthropic's Claude Opus 5.5 at 58 points, Claude Sonnet 5.5 at 56 points, and Muse Wonder Full at 56 points.

Moonshot plans to publish the model weights by November 4, 2026. If that release goes as planned, K3.1 Pro will become the highest-scoring open-weights model available. The gap between open and proprietary models would shrink from 12 points to 3 points. This assumes no other model changes occur in the interim.

The company also offers K3.1 Pro Highspeed. This variant runs two to three times faster than the standard version. Both options are available through Kimi.ai, Kimi Work, Kimi Code, and the Kimi API.

Performance gains were most visible in agentic tasks. On the GDPval-AA benchmark, K3.1 Pro reached 1,846 Elo. That is 312 points above K3. It trails only Claude Opus 5.5 at 1,866 and Muse Wonder Full at 1,857. On Terminal-Bench 4.0, Moonshot reports a score of 65.7 percent. This beats GPT-6 Astra at 58.2 percent. It is slightly below Claude Opus 5.5 at 66.4 percent. The previous K3 model managed only 13 percent on that test.

Hallucination rates dropped sharply. On the internal Kimi Hallu Bench Hard, claim-level hallucinations fell from 23.4 percent with K3 to 8.3 percent. Moonshot describes this as a stress test rather than a general error measure. GPT-6 Astra still leads this specific metric at 5.5 percent. Finance benchmarks also improved. The model gained 7.2 points on Financial Modeling and 11.6 points on Financial Research. Moonshot states it tops all tested models on OfficeQA Pro v2.

The architecture remains unchanged from K3. It features 2.8 trillion total parameters with 104 billion active. The context window holds one million tokens. API prices dropped by 33 percent. Input costs are $2 per million tokens. Output costs are $10 per million tokens. Cache hits cost $0.20 per million tokens. Because the model generates 14 percent more output tokens, the cost per Intelligence Index task is $1.98. This is nearly identical to the $2.00 cost for K3.